“Psyosphere” : A GPS Data Analysing Tool for the Behavioural Sciences

“psyosphere”

A GPS Data Analysing Tool for the Behavioural Sciences Benjamin Ziepert

Master Psychology of Conflict, Risk and Safety University of Twente

1

Supervisor University of Twente: Dr. Ir. Peter W. de Vries

2

Supervisor University of Twente: Dr. Elze G. Ufkes

Abstract

Positioning technologies (PT) such as GPS are widespread in society but are used only sparingly in behavioural science research. The current study attempts to unlock PT potential for behavioural science studies by developing a research tool to analyse GPS tracks, and by giving an overview of behavioural variables that can be studied with PTs. To test the research tool and to find more links between behavioural variables and PTs, we conducted two similar experiments. During the experiments, participants were placed in teams and carried cards with either a hostile or non-hostile task from a start to finish area. At the finish area the participants had to avoid guards, in order that their cards would not be confiscated. After each of three rounds the participants filled out a questionnaire to measure mental states related to hostile intent. The results show that the participants collectively changed their strategies on how to avoid guards, with each consecutive Round, and that mental states, such as fear, can be linked to changes in GPS variables, such as walking closer together. The current study demonstrates that behavioural experiments can be performed with GPS, outside of a laboratory setting.

Keywords: GPS, GIS, R, spatial movement, walking, psychology, hostile intent

Acknowledgments

First of all, I want to thank my parents for their many years of support. Since my early childhood my parents always encouraged my academic interests and none of my many projects would have been possible without them, including this thesis.

Further, I would like to thank Dr. Peter W. de Vries and Dr. Elze G. Ufkes for their never-ending and extraordinary support for my thesis. Peter’s and Elze’s patience and feedback allowed me to greatly improve my academic skills and to discover new and exciting statistical methods.

Especially Peter provided me again and again with opportunities to give lectures about what I have learned for this study, let me supervise students that used my research tools for their thesis, and let me present my research to other researchers. I am very thankful for all these great opportunities.

I also want to thank the head of the department Prof. Dr. Ellen Giebels. Ellen asked me as a bachelor student to help develop an experiment for all students of an undergraduate course and to create a research tool to analyse the data. The experiment and research tool created the basis for this thesis. Later, Ellen included me in a research project for the Dutch TV show Hunted and allowed me to test my new data analysis tool “psyosphere” as part of the project.

Moreover, I would like to thank my colleagues at the department Conflict, Risk and

Safety at the University of Twente for the many teaching opportunities they provided. By

giving lectures and supervising students I was able to share my knowledge that I gained for

this study.

I’m very grateful for my girlfriend Çiğdem. She motivated me every day to stay focused, gave me valuable feedback, and she was supportive even when I had to work on our free days.

Finally, I want to thank my many friends for their input and feedback. Especially

Jochem Goldberg, Melle Koedijk and Evelien Geertsema provided me with valuable insights.

“psyosphere”

A GPS Data Analysing Tool for the Behavioural Sciences

Positioning technologies (PTs) such as Global Positioning Systems (GPS), Glonass and Galileo can be used to determine the position on the globe and to record for instance the movement of planes, cars, and individuals (Hofmann-Wellenhof, Lichtenegger, & Wasle, 2007). PTs are now omnipresent in mobile devices such as smart phones, tablets, and laptops.

It could for instance be used to identify people with early warnings signs for depression (Palmius et al., 2017; Saeb et al., 2015), partly or fully replace self-reported diaries in traffic research (Bohte & Maat, 2009; Schuessler & Axhausen, 2009; Stopher, Bullock, & Horst, 2002;

J. Wolf, 2006), determine how populations behave after a disaster such as an earth quake (Bengtsson, Lu, Thorson, Garfield, & Von Schreeb, 2011), or to automatically detect active pickpockets in a shopping mall (Bouma et al., 2014). This omnipresence makes PTs potentially interesting to study behaviour in naturalistic settings. Surprisingly, behavioural scientists use PTs only to a small extent.

In this paper we argue that there are two reasons why PTs have largely been neglected in behavioural research. First, the data are too complex to analyse with software that traditionally were used in the social sciences, such as IBM SPSS Statistics (SPSS). Second, only a limited number of studies investigated the relationship between psychological variables and PT data. Consequently, little information is available which psychological variables could be studied with PTs. Therefore, the aim of this study is to develop a tool that enables behavioural scientists to make readily use of PTs to study movement and to give an overview of psychological variables that can be studied with PTs.

In current behavioural science research, the assessment of movement is often done

via trained observers, interviewers, or self-reported diaries, and these methods have been of

great use in the past (Axhausen, Zimmermann, Schönfelder, Rindsfüser, & Haupt, 2002;

Doherty & Miller, 2000; Goodchild & Janelle, 1984; Janelle, Goodchild, & Klinkenberg, 1988;

Shoval et al., 2010). These conventional methods to measure movement come with drawbacks that may be circumvented by using PTs instead. According to Shoval et al. (2010), the main obstacle is the information provided by participants. For example, people frequently underreport trips that are small, and people also underreport trips that do not start or end at home. Moreover, participants that drive a car underestimate their travel time whereas public transportation users overestimate their travel time (Ettema, Timmermans, & van Veghel, 1996; Stopher, 1992). Furthermore, participants can consciously omit information, for instance, if answers or not socially desirable. Finally, the interviewer could fail to prompt recall (interviewer error), or the participants could simply forget the information over time (recall bias; Anderson, 1971; Golledge, 1997; Vazquez-Prokopec et al., 2009). These limitations can be compensated by using PTs such as GPS (Bohte & Maat, 2009).

Benefits of PTs

Especially in traffic research, scientists compared PTs such as GPS with traditional methods of movement tracking and they pointed out several benefits of using PTs (Bohte &

Maat, 2009; Schuessler & Axhausen, 2009; Stopher et al., 2002; J. Wolf, Schönfelder, Samaga,

Oliveira, & Axhausen, 2004). Compared to self-reported diaries or interviews, (1) GPS loggers

are less intrusive, as loggers may substantially reduce information that needs to be self-

reported by participants or need to be asked by interviewers. (2) GPS loggers can reduce costs

by reducing the interview duration. (3) The survey periods can be longer; smartphone apps

tracking movement in the background allow for longer data-collection periods compared to

when the participants self-report their trips. (4) The data quality can be improved since GPS

loggers report small trips and travel times more accurately. (5) Finally, the sensors also have

the benefit of recording additional data such as speed and acceleration which can be used for additional analysis (J. Wolf et al., 2004).

Next to these examples from traffic research, there are studies in other areas that employed PTs to replace or augment traditional methods of movement tracking. Particularly, research with target groups that are unable to maintain a self-reported diary and where observers would be especially expensive. For instance, for the mentally impaired, children and the elderly it may be difficult or even impossible to maintain a diary (Shoval et al., 2011).

Traditionally, caretakers or family members were used to monitor those participants and noted the activities or filled in behavioural checklists for them (Shoval et al., 2011). Using caretakers or family members can be quite expensive, burdensome and biased. Moreover, Isaacson, Shoval, Wahl, Oswald, and Auslander (2016) argue that researchers may even avoid doing experiments with these target groups at all, because of these obstacles.

For groups that cannot maintain a diary, PTs such as GPS loggers can be an option to replace observers (Isaacson et al., 2016; Shoval et al., 2010; Shoval et al., 2008; Shoval et al., 2011). A critic could wonder whether a participant who is unable to fill in a diary would be able to handle the complex protocol for using sensors. Fortunately, research has shown that the mentally impaired and the elderly are indeed able to follow these protocols (Isaacson et al., 2016).

Additionally, as with many digital technologies, digital position recognition has some

strengths compared to analogue data gathering (Brynjolfsson & McAfee, 2014). First of all,

the analysis can be automated. For instance, an algorithm to detect pickpockets (Bouma et

al., 2014) can be used again and again to detect this behaviour without the intervention of a

researcher. Second, if the sensors are directly connected to a processor, the analysis can be

real-time. The Global System for Mobile Communications (GSM) or Wi-Fi can often be directly

connected to a processor, but this is not always possible with GPS loggers. Third, the analysis can be scaled up relatively easily. Therefore, it is possible to use the pickpocket classification algorithm on a larger airport by buying more sensors, for a fraction of the costs necessary to hire and train more security personnel. Fourth and finally, the analysis can be transferred easily. Once the technology is developed, it can be used on separate locations with a comparable small investment cost. For instance, installing new hardware and sensors can be cheaper than hiring and training new observers for a new location.

PT usage in past research

As mentioned before, PTs can be utilized to study a variety of subjects. For instance, research has shown that measures such as positive affect, extraversion or openness to experiences can predict the number of places someone visits over several days (Byrne &

Byrne, 1993; Schwerdtfeger, Eberhardt, Chmitorz, & Schaller, 2010; P. S. A. Wolf, Figueredo,

& Jacobs, 2013). Another example is risk-taking behaviour. GPS loggers can be used to detect risky driving behaviour such as speeding (Bolderdijk, Knockaert, Steg, & Verhoef, 2011).

Table 1 gives a broad overview of research that employed PTs to study behaviour.

Table 1

PTs and Their Use in Past Research

Measures Research

Anxiety, depression, or lifestyle (e.g. positive affect or extraversion)

Determining relationship between active versus sedentary lifestyle, social anxiety and depression, and number places visited with GPS (Huang et al., 2016; Saeb, Lattie, Schueller, Kording, & Mohr, 2016; P. S. A. Wolf et al., 2013).

Community specific routes description and

visualisation

Measuring segregation in city communities with GPS (Davies et al., 2017;

Whyatt et al., 2017).

Depression detection Detecting depression from GPS movement data characteristics such as location variance, home stay, or mobility between favourite locations (Palmius et al., 2017; Saeb et al., 2015).

Environmental exposure Measuring daily environmental exposure with GPS (Chaix et al., 2013;

Phillips, Hall, Esmen, Lynch, & Johnson, 2001).

Following and leadership detection

Detecting leadership and followership with movement patterns (e.g. co- moving) with Wi-Fi data. (Kjargaard et al., 2013).

Information or disease spreading characteristics

Studying information spreading in face-to-face networks with Bluetooth, RFID and Wi-Fi (Isella, Romano, et al., 2011; Isella, Stehlé, et al., 2011;

Madan, Moturu, Lazer, & Pentland, 2010).

Physical activity Measuring physical activity of children, the elderly or other target groups with GPS (Elgethun, Fenske, Yost, & Palcisko, 2002; Fjørtoft, Kristoffersen, &

Sageie, 2009; Isaacson, D’Ambrosio, Samanta, & Coughlin, 2015; Krenn, Titze, Oja, Jones, & Ogilvie, 2011; Maddison & Ni Mhurchu, 2009; Shoval et al., 2011).

Pickpocket detection Detecting pickpockets with movement characteristics (e.g. walking speed) measured with security cameras (Bouma et al., 2014).

Population movement characteristics

Studying population behaviour after a disaster with GSM (Bengtsson et al., 2011).

Risk seeking Measuring speeding as a form of risk seeking with GPS (Bolderdijk et al., 2011).

Travel characteristics such as travel mode, route choice or speed

Studying travel behaviour such as travel mode, route choice or speed with GPS (Bohte & Maat, 2009; Draijer, Kalfs, & Perdok, 2000; Murakami &

Wagner, 1999; Necula, 2015; Schuessler & Axhausen, 2009; Stopher et al., 2002; J. Wolf, 2000, 2006; J. Wolf et al., 2004).

Virus transmission risk Studying the spreading of disease with GPS (Vazquez-Prokopec et al., 2013;

Vazquez-Prokopec et al., 2009).

Walking routes Assessing tourist walking routes with GSM and GPS (Xia, Arrowsmith,

Jackson, & Cartwright, 2008).

As can be seen in Table 1, there are only a small number of studies investigating the link between The past research (see Table 1) contains only a small number of studies that investigated the link between PT data and psychological variables, such as personality or mental states (e.g., Palmius et al., 2017; Saeb et al., 2015). Therefore, we want to investigate if more psychological variables, than mentioned in Table 1, can be linked to PT data.

Laboratory studies have shown that behaviour may become overt as a result of psychological variables. For instance, sad, depressed and frightened people tend to walk slower than others, and joy and anger are linked to increased walking speed (Barliya, Omlor, Giese, Berthoz, & Flash, 2012; Gross, Crane, & Fredrickson, 2012; Michalak et al., 2009). Other research indicates that personality traits such as agreeableness are also linked to increased walking speed (Satchell et al., 2017).

Hostile intent and movement

Research outside of the laboratory has shown that motivation or conscious decisions such as pickpocketing corresponds with specific body movement (Bouma et al., 2014). Their algorithms to detect pickpockets based on variations in walking speed, orientation change or distance to other people were shown have a sensitivity up to 95.6% with 0.5% false alarms.

Researchers argue that other behaviours such as smuggling can also result in behavioural changes that can be detected (Wijn, Kleij, Kallen, Stekkinger, & de Vries, 2017).

They conducted an experiment where the participants transported packages with supposedly

illegal and legal contents. Participants were recorded on video and independent lay observers

were asked to watch the videos and rate which participants were transporting an illegal

package. According to the researchers the mental processes while transporting an illegal

package lead to changes in the participants’ behaviour that could be detected by the

observers. However, the researchers did not discuss which cues could be used for the

detection and further research is needed. Therefore, the current study will investigate if the mental processes can be linked to measurable changes in movement.

The mental processes of transporting an illegal package are linked to hostile intent.

Wijn et al. (2017) define hostile intent “as an individual’s intent to act in ways that imply or aim to inflict harm onto others.” (p. 2). People with hostile intent try to hide it when they expect that others will try to prevent their actions (DePaulo et al., 2003; Ekman, Friesen, &

O'sullivan, 1988; Koller, Wetter, & Hofer, 2016; Wijn et al., 2017).

Wijn et al. (2017) argue that persons with hostile intent have a heightened state of self-saliency and interpret cues in the surroundings as to be connected to them. A cue could, for instance, be a police guard looking in the direction of that person. This cue can cause a fear-related response pattern (e.g. fight or flight) and the person will try to supress the fear response in order not to attract the attention of the guard. In other words, a person with hostile intent will try to act normal. This suppression of fear-related responses is a cognitive effortful process and can be constrained, by other cognitive tasks (e.g. counting), or fatigue.

Therefore, people with hostile intent should show more deviant behaviour if they have an increased cognitive load and get cues from the environment that they perceive as related to them (Wijn et al., 2017).

As an example, it could be argued that an unexpected route change by construction works at an airport could increase the cognitive load. When someone needs to reorient him- or herself in an unfamiliar environment it increases the cognitive load and therefore limits the person’s ability to suppress fear related responses. Additionally, more security guards could act as a cue to trigger detectable behaviour changes (Wijn et al., 2017).

A PT research tool for behavioural scientists. As mentioned before, we argue that

data from PTs are unsuitable to be analysed with the software that is traditionally used by

behavioural scientists such as SPSS. SPSS is a specialised software to perform statistical analysis which is not suitable to handle geospatial data and analysis. Therefore, we developed a data analysis software with the aim to enable behavioural scientist to analyse movement data without the need of additional special expertise.

Out of the variety of PTs we decided to focus on GPS for our tool. GPS can be used all over the globe and does not depend on local GSM, Wi-Fi or other infrastructure. GPS is also omnipresent in smart phones or other devices, and dedicated GPS loggers are affordable. The data analysis software will work with longitude, latitude and timestamp data points that are typical for GPS loggers. The movement data from other PTs such as GSM and Wi-Fi data can be converted to be used with the same software once it is converted to longitude and latitude.

The tool will be a R package and is called “psyosphere” (Ziepert, Ufkes, & de Vries, 2018; see Appendix 3). R is an open-source programming language and data-analysis tool that is becoming more widespread (Muenchen, 2012). The choice for R has several benefits: since R is used by psychologists and computer scientists it could improve cooperation of interdisciplinary teams, the software is free of charge and therefore easier accessible than for instance SPSS, there are pre-existing packages for spherical calculations and handling of GPS data (e.g. Hijmans, Williams, & Vennes, 2015; Kahle & Wickham, 2013; Loecher & Ropkins, 2015; Wickham, 2016), and since R is open source the software can be improved upon by the research community.

The current study

For our current study we will focus on movement-descriptive variables such as speed,

distance to peers or variations in the route. As mentioned before, there are several

psychological variables that are linked to movement. For example, emotions such as

happiness or fear are linked to walking speed (Barliya et al., 2012; Gross et al., 2012; Michalak

et al., 2009). To find more psychological variables that are linked to movement, we want to determine whether mental states related to hostile intent result in changes in descriptive movement variables measured with GPS. Two experiments were conducted that were part of an undergraduate course for psychology students. During the experiments the participants would wear GPS loggers and were to transport cards with supposedly legal and illegal tasks from a start to finish area. In the finish area the participants had to avoid guards that could confiscate the cards. After each Round, the participants would fill in a questionnaire to measure the mental states.

Methods Participants

We conducted two experiments as part of an undergraduate psychology course at the University of Twente. The first experiment took place in 2014 and the second in 2015. The experiments are similar to each other and we did not conduct an analysis between the experiments. In the first experiment 64 students participated, two were excluded from the analysis due to sensor failure, and 62 students (44 female and 18 male) remained. The average age was 21.61 (SD = 5.60) and ranged from 18 to 37. Furthermore, 30 students were Dutch and 32 were German. In the second experiment 93 students participated, 19 students were excluded from the analysis due to a lack of sensors or sensor failure, and 74 students (51 female and 23 male) remained. The average age was 22.41 (SD = 5.60) and ranged from 18 to 46. Moreover, 38 students were Dutch and 34 were German. The participants that acted as guards were excluded from analysis to limit the scope of the current study.

Procedure

The participants signed up for the experiment during an introductory lecture for an

undergraduate course. The experiment was explained to the participants and they received

written instructions. They had the possibility to ask questions and afterwards signed an informed consent form. The participants were randomly assigned into teams of five to six people or got an individual task. The participants received a GPS logger and were told to gather three hours later in a small park on the university campus.

Tasks. The teams (smugglers) had the tasks to transport supposed contraband during the experiment and the individual assignment for the other participants was to find the participants with contraband (guards). The contraband was a paper card with the size of a card game card. On the card was either an image of cocaine (illegal card) or flour (legal card) printed and with a text indicating the same. Both teams and guards could gain points by fulfilling their task and it was announced that the best team and best guard would win a price.

The price was a voucher for a cinema and a bar of chocolate for each winning team member and guard.

Area. After arriving at the park, participants were directed to their assigned locations.

The teams would go to a starting point that was behind a mount and out of sight of the guards.

The guards were waiting at the finish in an about two-metre-wide and 20-metre-wide strip

that had to be crossed by the teams later. The finish area was marked with barrier tape on

the ground. A group of 17 tall trees were standing inside and around the finish area. A visual

inspection of the data did not reveal signal distortions by the trees. The distance from the

starting area to the finish area was 150 metres. The mound was in the middle between start

and the finish and the teams and guards could see each other when the teams walked over

the mound. The mound had a semi-circular shape and the guards were positioned in the

centre of the semi-circle in a distance of about 75 metres. See Figure 1 for an illustration of

the area.

Figure 1. Experiment area with participant tracks and GPS polygons. The tracks of six team

members in Experiment 2 are plotted in black (illegal card) and yellow (legal card). The tracks begin in the start polygon (A), enter the line of sight for guards’ polygon (B) and end before the finish polygon (C).

Contraband. The teams would receive the legal and illegal cards at the starting area, they had to distribute it and each participant had to carry exactly one. The cards stated that the teams would win ten points for each illegal card they transported and one point for each legal card. The legal card also stated that the guards would lose a point if they took a legal card from a team member. Before starting the teams had to write a number on the card that was matched to their GPS logger and the starting time.

A B C

Round. The teams were instructed to walk from the starting area, across the park and through the finish area. The guards could confiscate the team members card by tapping them on their shoulder. The team member would give their card to the guard and the guard would note a number on the card which was assigned to the guard. The team members had to avoid being checked by the guards. This could be done by for instance distracting them by sending the team members with legal cards first or walking with a wide distance among each other therefore it would be difficult for the guards to reach all team members before they crossed the finish area. The guards were not allowed to leave the finish area and had to wear safety vests to enable the team members to spot them easily. Each time after crossing the finish area the team members would drop the remaining cards they had into a box and fill in a questionnaire. After this they would walk back to the start position for another Round.

Experiment 1. Between the experiments were some differences on how the

experiments were conducted. In Experiment 1, five participants had the task to be a guard

and the other participants were assigned into teams between three and six members with an

average of 4.85. Before the start of the first experiment the participants filled in an additional

trait questionnaire. Further, participants were instructed not to run, and four rounds were

conducted. Additionally, all teams were wearing a card with their team number on their chest,

and a team of five participants was wearing stress sensors around their wrist that measured

their skin conductivity. At the starting area each team got two illegal cards and the rest were

legal cards. Afterwards, between four and five teams with an average of 21.08 participants

would start at the same time and the ratio between guard’s participants was 0.24 when the

participants approached the finish area. After each Round the teams would fill in a Dutch

version of the State questionnaire.

Experiment 2. In Experiment 2, three participants had the task to be guards and the other participants were assigned into teams between four and seven team members with an average of 5.69. Furthermore, the instruction not to run was omitted, and the participants were not wearing any stress sensors or team numbers. At the starting area the teams could choose freely the ratio of illegal and legal cards and they were asked to write down which strategy they wanted to use. Afterwards, each team would start separately, and when approaching the finish area, the ratio between participants and guards was 0.54. Additionally, the finish area in Experiment 2 was larger than in Experiment 1 and enabled the guards to walk more freely. At the end of each Round, the team members would write down their points and could see the total points of the other teams. Finally, they would only fill in an English version of the State questionnaire. Some questions were removed and added in the State questionnaire (see Appendix 1) and additionally the teams were asked to rank how they perceived every team member as a leader and how they could improve their strategy as a team in the next Round. The guards were frequently told how many points each guard had.

Measures

State questionnaire. The mental states of the participants were measured with a

questionnaire based on the research by Wijn et al. (2017) and Stekkinger (2012) to measure

hostile intent and related constructs. Some questions have changed to fit the current study

and two questions were added to measure self-observed behaviour changes. For instance,

whether the participants changed their pace after seeing the guards. Table 2 contains all

questions, and the Cronbach’s alpha or Pearson's R for the state questionnaire constructs

when applicable. The reliability for each scale where we calculated Cronbach’s alpha, the

index was above .78 (see Table 2). All State questions used a 7-point Likert scale from 1 “Not

at all” to 7 “Very much”. Appendix 1 contains the full questionnaire.

Two items checked whether participants felt that they had hostile intentions (Hostile Intent; Stekkinger, 2012; Wijn et al., 2017). Five items measured the participants alertness to threats from the guards (Alertness to Being Target of Guards; Galbraith, Manktelow, & Morris, 2008; Stekkinger, 2012; Wijn et al., 2017). Five items checked the inhibitory and activation control (Cognitive Self-Regulation; Stekkinger, 2012; Wijn et al., 2017). Four items measured the self-focus of the participants (Situational Self Awareness; Govern & Marsch, 2001;

Stekkinger, 2012; Wijn et al., 2017). Four items assessed the feelings of fright that the participants felt through the presence of the guards (Frightened by Presence of Guards;

Stekkinger, 2012; Wijn et al., 2017). Five items checked the impulses that were suppressed by the participants (Suppressed Impulses to Change Movement; Stekkinger, 2012). Three items measured the extent that participants questioned themselves (Contemplation of Hostile Intent; Stekkinger, 2012; Wijn et al., 2017). Finally, two items are added to the questionnaire and assessed the self-observed behaviour changes (Awareness Movement Change in Presence of Guards; Stekkinger, 2012). For a detailed explanation of the mental processes and their function see Wijn et al. (2017).

GPS data. Every participant carried an i-gotU GT-600 GPS logger. The logger received location signals from GPS satellites and saved them every second in a data point. The logger saved the time, latitude, longitude and elevation. From the GPS data Speed, Speed Variation, Intra-Team Distance, Route Deviation and Variation Route Deviation were calculated. Speed was measured as the mean kilometres per hour between each data point. Speed Variation was calculated as the standard deviation of the kilometres per hour between each data point.

Intra-Team Distance is the mean distance from one team member to the other team members

in metres. Route Deviation is the distance in metres between a data point and the shortest

route from start to finish. Variation Route Deviation is the standard deviation of the Route Deviation for each data point.

Analysis

GPS data preparation. The data from the GPS loggers were exported and analysed with the R package “psyosphere”, which was developed for the current study (Ziepert et al., 2018). The software created a track for each Round of each participant, and plotted the tracks on a map, which was retrieved from google maps (Google LLC, 2018). See Figure 1 for the tracks and map. A track began in the starting area that was determined by a polygon of GPS coordinates (A) and ended when the participants crossed a GPS based finish line behind the finish area (B). The R package also marked automatically from which point the teams and guards could see each other based on a polygon of GPS coordinates (C). Before the point of visibility, the teams followed generally a straight line and started to change their movement mostly after seeing the guards. Therefore, the analysis included only the data from when the teams were visible to the guards, until the members crossed the finish line. Within line of sight of the guards 31,113 coordinates were recorded in Experiment 1, for four rounds, and 17,172 in Experiment 2, for three rounds. Based on this data, the R package calculated the GPS variables that are mentioned above.

The R package detected two types of faulty data. First, if the speed exceeded a

maximum of 40 km/h then the data were marked as missing values and excluded from

analysis. An unrealistic speed can be for instance recorded due to signal loss from the GPS

satellites. This occurred 16 times (0.05%) in Experiment 1 and 8 times (0.05%) in

Experiment 2. Second, if the time difference between coordinates exceeded one second then

the Speed, Speed Variation and Distance between the coordinates were marked as missing

values and excluded from analysis. Three coordinates (0.01%) in Experiment 1 and 152

coordinates (0.89%) in Experiment 2 were excluded because of a time difference larger than one second.

State PCA. We analysed the State questions that were used in both experiments with a principal component analysis (PCA). In total, we conducted six explorative PCAs, one for each of the three rounds in the two experiments. Afterwards, we compared the PCAs and counted how often items shared a component. A model with eight components emerged and we testes this model with a confirmatory PCA. For the confirmatory analysis, the data of the six rounds over the two experiments were analysed together.

Relationships between State and GPS variables. Descriptive statistics and correlations, for the State components and GPS variables, were calculated for each experiment separately. Finally, we conducted a multi-level analysis with the GPS variables as dependent variables, and the State components and rounds as the predictors. In total we created ten models, five for each experiment. The multi-level analysis tested for consistent changes per Round (e.g. increasing Intra-Team Distance per Round) and the impact of grouping in teams. Three random effect models did not converge, and two of these models, were models with a maximum random effects structure based on the experimental design.

Moreover, According to Barr, Levy, Scheepers, and Tily (2013), a maximum random effects

model should be prioritized when conducting a multilevel analysis. For our current study, a

maximum random effects model included random slopes per Round and a static intercept per

team and participant. To improve the model convergence rate, Barr et al. (2013) suggest to

remove outliers, and therefore, data have been removed from the GPS variables, except Intra-

Team Distance, when the data were outside of the Inter Quartile Range times 1.5 (Hoaglin,

2003). 15 outliers have been removed from Speed, 9 from Speed Variation, 5 from Route

Deviation, and 14 from Variation Route Deviation. After removing the outliers, all models with

a maximum random effects structure converged and two models did not converge. Intra- Team Distance was excluded from the outlier removal since this increased the model convergence rate.

Exclusions. Only the first three rounds of both experiments were used since the participants did not complete the State questionnaire after the fourth Round of Experiment 1.

The stress sensor data was not used due to faulty data; answers to strategy, leadership and motivation questions were not analysed to limit the scope of the current study and could be analysed in a follow up study.

Results Factor Analysis

We conducted an exploratory principal component analysis (PCA) for 30 items of the

state questionnaire (see Appendix 1) for each experiment and for each of the first three

rounds. This resulted in six PCAs and each PCA used an oblique rotation (oblimin). Afterwards

we compared the PCAs by counting how often items were grouped together within a

component. We used the resulting model with eight components for a confirmative PCA over

the data of both experiments and their first three rounds. The Kaiser-Meyer-Olkin (KMO)

measure for the confirmative PCA verified the sampling adequacy for the analysis. The KMO

of .90 is above the minimum of .50 (Kaiser, 1974). The Bartlett’s test of sphericity, χ²

(435) = 8198, p < .001, illustrated that the correlations between the items were large enough

for a PCA. The eight components explained 22% of the variance. Seven items had an

eigenvalue higher than 1.00 and the eighth component had an eigenvalue of 0.88. As already

mentioned, the model that we derived from the first six PCAs indicated eight components

and not seven. Therefore, we decided to retain eight components for the confirmative PCA

that was conducted for all data combined.

Table 2 shows the factor loadings after rotation in the pattern matrix. The coefficients

in the pattern matrix indicate the unique contribution of a component to an item while

controlling for other components. Table 3 shows the structure matrix of the PCA. The

coefficients in the structure matrix indicate the relationship strength between the item and

the component while ignoring other components. The clustering of the items show that the

items load on the components as intended. The constructs in order are Alertness to Being

Target of Guards, Cognitive Self-Regulation, Situational Self Awareness, Frightened by

Presence of Guards, Suppressed Impulses to Change Movement, Contemplation of Hostile

Intent, Awareness Movement Change in Presence of Guards and Hostile Intent.

Table 2

PCA Pattern Matrix for the State Questionnaire

Item Alertness to Being Target of Guards Cognitive Self-Regulation Situational Self Awareness Fright Suppressed Impulses to Change Movement Contemplation of Hostile Intent Awareness Movement Change in Presence of Guards Hostile Intent I thought I had attracted the border guards’ attention .89 -.02 .02 -.04 -.00 .01 -.01 .01 I had the feeling the border guard(s) targeted me .89 .04 -.02 -.03 .03 .03 -.01 -.07 I felt like I was the one being addressed by the border guard(s) .87 -.01 -.01 .01 .04 .03 -.02 .04

I had a feeling that I was going to be stopped .79 .07 .04 .05 .09 -.02 -.04 -.02

I had the idea that the others were paying attention to me .70 -.07 .04 .10 -.13 -.07 .19 .08 During this round I have tried to hide my nerves .07 .88 -.00 .09 -.03 -.04 .01 .03 During this round I have tried to hide my tension .06 .86 .03 .03 -.06 -.03 -.02 .04 During this round I have tried to hide my emotions .06 .84 -.01 .03 .02 .00 .03 .00 During this round I have tried not to attract attention -.13 .79 -.03 -.10 .05 .11 .04 -.06 During this round I have tried to act as normal as possible -.14 .62 .14 -.07 .09 -.09 .02 .08 During this round I was aware of the way I presented myself .00 -.02 .88 -.02 -.04 .02 .01 .05 During this round I was aware of how I looked -.02 .07 .86 .01 -.03 -.05 -.03 -.05 During this round I was aware of my inner feelings .05 .04 .77 .06 .00 .05 -.11 .07 During this round I was aware of everything in my direct surroundings .02 -.08 .74 -.02 .08 .03 .10 -.04

I was startled by the border guards’ presence .00 -.04 .07 .81 .05 -.14 .08 .10

The border guards’ presence made me feel stressed .06 .08 -.02 .79 .06 .14 -.11 .02 I was startled when I first noticed the border guards .03 .03 -.02 .79 .05 -.17 .10 .10

The border guards’ presence made me feel tense -.01 .08 .04 .77 .01 .27 -.00 -.09

I would rather have chosen a different route .19 .01 -.02 -.13 .77 .02 -.07 .06

I would rather have taken a detour to avoid the border guards .03 .02 .06 .02 .77 .10 .05 -.07

I would rather have hidden myself -.07 .03 .02 .15 .72 .11 .03 -.06

I would rather have turned around .02 -.00 -.06 .16 .71 -.08 .00 .14

I would rather have run away from the border guards .02 .02 .05 .10 .52 -.23 .24 .20 I was thinking about what I had to hide from the border guards .02 .03 .12 -.03 .07 .75 .14 .13 I was wondering whether I looked suspicious to the border guards .05 .07 .08 .22 .07 .61 .14 -.12 I was wondering whether I was doing something that I was not allowed to do .03 .03 .02 -.00 .07 .60 .01 .43 During this round I have increased my pace as soon as I saw the border guards -.00 .03 -.06 .05 -.06 .05 .85 .13 During this round I have changed my course as soon as I saw the border guards .04 .06 .08 -.03 .13 .07 .78 -.16 During this round I felt I was doing something illegal .02 .15 .04 .06 .09 .14 -.06 .74

During this round I felt I had hostile intentions .02 .01 .09 .11 .03 .01 .12 .73

Eigenvalues 3.70 3.52 2.92 3.08 2.96 1.79 1.74 1.77

% of variance .12 .12 .10 .10 .10 .06 .06 .06

α (R) .90 .87 .84 .88 .83 .78 (.53) (.64)

Note. Component loadings that are higher than or equal to .40 are in bold. Data of all

experiments and rounds are analysed together.

Table 3

PCA Structure Matrix for the State Questionnaire

Item Alertness to Being Target of Guards Cognitive Self-Regulation Situational Self Awareness Frightened by Presence of Guards Suppressed Impulses to Change Movement Contemplation of Hostile Intent Awareness Movement Change in Presence of Guards Hostile Intent I felt like I was the one being addressed by the border guard(s) .89 .13 .19 .26 .38 .10 .11 .20 I thought I had attracted the border guards’ attention .88 .11 .20 .20 .32 .07 .09 .15

I had the feeling the border guard(s) targeted me .88 .15 .17 .20 .35 .09 .10 .09

I had a feeling that I was going to be stopped .85 .22 .24 .31 .42 .09 .12 .18

I had the idea that the others were paying attention to me .71 .08 .21 .30 .22 -.00 .28 .23

During this round I have tried to hide my nerves .21 .91 .34 .41 .26 .22 .20 .27

During this round I have tried to hide my tension .18 .87 .35 .34 .20 .21 .15 .25

During this round I have tried to hide my emotions .20 .87 .32 .36 .28 .25 .20 .24 During this round I have tried not to attract attention -.03 .77 .24 .17 .18 .31 .15 .10 During this round I have tried to act as normal as possible .01 .65 .33 .20 .20 .14 .15 .23 During this round I was aware of the way I presented myself .18 .30 .88 .17 .12 .26 .20 .23

During this round I was aware of how I looked .15 .34 .85 .16 .09 .18 .15 .15

During this round I was aware of my inner feelings .23 .35 .81 .24 .18 .28 .12 .27 During this round I was aware of everything in my direct surroundings .20 .22 .75 .17 .21 .24 .27 .16

I was startled by the border guards’ presence .27 .29 .24 .88 .40 .02 .36 .41

I was startled when I first noticed the border guards .28 .31 .17 .87 .39 -.02 .36 .41

The border guards’ presence made me feel stressed .30 .40 .22 .85 .43 .26 .20 .36

The border guards’ presence made me feel tense .22 .42 .28 .81 .38 .39 .29 .27

I would rather have taken a detour to avoid the border guards .34 .27 .22 .35 .82 .31 .29 .19

I would rather have chosen a different route .45 .19 .13 .23 .79 .19 .14 .24

I would rather have turned around .33 .23 .10 .47 .79 .11 .25 .36

I would rather have hidden myself .25 .29 .19 .43 .78 .31 .28 .21

I would rather have run away from the border guards .31 .24 .20 .45 .65 .00 .43 .40 I was thinking about what I had to hide from the border guards .16 .35 .41 .23 .35 .85 .34 .32 I was wondering whether I looked suspicious to the border guards .21 .37 .35 .38 .36 .70 .34 .15 I was wondering whether I was doing something that I was not allowed to do .18 .33 .32 .29 .35 .70 .22 .57 During this round I have increased my pace as soon as I saw the border guards .11 .21 .18 .34 .23 .20 .87 .29 During this round I have changed my course as soon as I saw the border guards .18 .24 .29 .27 .35 .25 .82 .07 During this round I felt I was doing something illegal .23 .41 .31 .43 .38 .32 .18 .84

During this round I felt I had hostile intentions .23 .29 .33 .46 .33 .20 .33 .83

Note. Component loadings that are higher than or equal to .40 are in bold. Data of all

experiments and rounds are analysed together.

Descriptive Statistics

Table 4 displays preliminary descriptive statistics such as the mean, standard deviation

and correlation coefficients for the State and GPS variables per experiment. For example,

more participants carried an illegal card (Illegal Card Selection) in Experiment 2 than in

Experiment 1 (M

= 0.41, M

= 0.88). In Experiment 1 each team had only two illegal cards to distribute, and in Experiment 2 the teams could choose a free ratio of cards and chose on average more illegal cards than in Experiment 1. Also, Hostile Intent was higher in Experiment 2 than in Experiment 1 (M

= 2.36 M

= 3.02). To determine the relationship between Illegal Card Selection and Hostile Intent an independent t test was conducted. The test results show that there was a significant difference in Hostile Intent for Experiment 1 when the illegal card was chosen (M

= 3.07, SD

= 1.87) than when the legal card was chosen (M

= 1.88, SD

= 1.07) with t

(112) = -5.11, p

< .001, and Cohen’s d

= 0.83. Similarly, there was also a significant difference in Hostile Intent for Experiment 2 when the illegal card was chosen (M

= 3.10, SD

= 1.64) than when the legal card was chosen (M

= 2.46, SD

= 1.31) with t

(49.04) = -2.60, p

= .012, and Cohen’s d

= 0.41. The results of both experiments demonstrate that participants who chose an illegal card reported a higher feeling of hostile intent than participants who chose a legal card, and this relationship was stronger in Experiment 1 than in Experiment 2. A reason for the stronger relationship in Experiment 1 could be that, in Experiment 1, each team had fewer illegal cards to distribute. The illegal cards scored ten points for the team and the legal cards scored one point, therefore, the illegal cards were important to become the team with the highest score. Since the illegal cards were limited in Experiment 1, it was important to carry the illegal cards successfully, that is, without being checked by the guards, to achieve the highest score. By comparison, in Experiment 2, there were unlimited illegal cards which created overall higher self-reported feelings of Hostile Intent but put less stress on the individual that carried an illegal card.

Differences between experiments. The descriptive statistics illustrate some

differences between the experiments. For instance, in the first experiment, participants were

warned not to run to prevent them from harming themselves. The warning was not given in

the second experiment. Consequently, in Experiment 1 the Speed was 4.59 kilometres per hour and 6.01 kilometres per hour in Experiment 2. Additionally, the Speed Variation was 1.43 in Experiment 1 and 4.04 in Experiment 2. It appears that, the warning not to run in Experiment 1 led to a decreased Speed and Speed Variation, compared to Experiment 2.

Other difference between the experiments are the Intra-Team Distance, Route Deviation, and Variation Route Deviation. During Experiment 1 the participants walked closer together than in Experiment 2 (M

= 9.59, M

= 12.65). Additionally, the Route Deviation was larger in Experiment 2 than in Experiment 1 (M

= 7.53, M

= 11.39), and also, the Variation Route Deviation was larger in Experiment 2 (M

= 3.15, M

= 4.77). A likely explanation for the differences, in Intra-Team Distance, Route Deviation, and Variation Route Deviation, could be that the participants in Experiment 2 had to avoid the guards more than in Experiment 1, since there were more guards per participant at the finish area in Experiment 2. Specifically, the ratio between guards and participants was 0.24 in Experiment 1 and 0.54 in Experiment 2.

Correlations. Table 4 illustrates that the State variables frequently correlate with each

other significantly, and the GPS variables correlate only partly with each other or with the

State variables. GPS variables that correlate with each other in both experiments are for

instance Variation Route Deviation and Intra-Team Distance (R

= .19, p

= .012, R

= .15,

p

= .030), and this means that teams which varied more in their route also had a longer

distance to their team members. A reason could be that some participants avoided the guards

by changing their route, other team members did not do so, and therefore the distance

between the participants increased. Additionally, Speed and Speed Variation (R

= -.48,

p

< .001, R

= .53, p

< .001), and Route Deviation and Variation Route Deviation (R

= .58,

p

< .001, R

= .62, p

< .001) correlated with each other. In both pairs, the relationships could

be expected since they are caused by the underlying mathematical relationship of the

variables. Specifically, the variables are the average and standard deviation of the same

measurement. For instance, a continuously increasing distance to the shortest path leads to

a higher average distance (Route Deviation) and a higher standard deviation of the distance

(Variation Route Deviation).

Table 4 1

Mean, SD and Correlation Coefficients of State Variables 2

Experiment Variable Mean SD 01

R (p) 02 R (p)

03 R (p)

04 R (p)

05 R (p)

06 R (p)

07 R (p)

08 R (p)

09 R (p)

10 R (p)

11 R (p)

12 R (p)

13 R (p)

14 R (p)

1 01 Illegal Card Selection 0.41 0.49 1.00

2 01 Illegal Card Selection 0.88 0.33 1.00

1 02 Hostile Intent 2.36 1.56 .38 (< .001) 1.00

2 02 Hostile Intent 3.02 1.61 .13 (.028) 1.00

1 03 Alertness to Being Target of Guards 3.71 1.72 -.01 (.874) .21 (.003) 1.00 2 03 Alertness to Being Target of Guards 4.05 1.81 .08 (.209) .30 (< .001) 1.00 1 04 Cognitive Self-Regulation 3.50 1.75 .36 (< .001) .41 (< .001) .11 (.126) 1.00 2 04 Cognitive Self-Regulation 3.91 1.39 .24 (< .001) .39 (< .001) .19 (.001) 1.00 1 05 Situational Self Awareness 4.07 1.59 .08 (.248) .36 (< .001) .24 (.001) .45 (< .001) 1.00 2 05 Situational Self Awareness 4.01 1.36 -.06 (.358) .41 (< .001) .28 (< .001) .35 (< .001) 1.00 1 06 Frightened by Presence of Guards 2.73 1.42 .33 (< .001) .62 (< .001) .23 (.003) .48 (< .001) .29 (< .001) 1.00 2 06 Frightened by Presence of Guards 3.32 1.55 .16 (.010) .48 (< .001) .39 (< .001) .38 (< .001) .28 (< .001) 1.00 1 07 Suppressed Impulses to Change Movement 2.76 1.45 .21 (.004) .65 (< .001) .33 (< .001) .44 (< .001) .33 (< .001) .65 (< .001) 1.00 2 07 Suppressed Impulses to Change Movement 2.96 1.46 .06 (.329) .36 (< .001) .55 (< .001) .22 (< .001) .17 (.005) .50 (< .001) 1.00 1 08 Contemplation of Hostile Intent 3.40 1.50 .16 (.023) .49 (< .001) .16 (.026) .47 (< .001) .54 (< .001) .46 (< .001) .55 (< .001) 1.00 2 08 Contemplation of Hostile Intent 3.32 1.54 .01 (.917) .55 (< .001) .31 (< .001) .43 (< .001) .37 (< .001) .41 (< .001) .41 (< .001) 1.00 1 09 Awareness Movement Change in Presence of Guards 3.26 1.82 .04 (.612) .32 (< .001) .01 (.875) .30 (< .001) .32 (< .001) .40 (< .001) .37 (< .001) .40 (< .001) 1.00 2 09 Awareness Movement Change in Presence of Guards 3.49 1.61 .05 (.368) .33 (< .001) .34 (< .001) .23 (< .001) .25 (< .001) .40 (< .001) .43 (< .001) .43 (< .001) 1.00 1 10 Speed 4.59 0.43 -.10 (.169) .06 (.468) -.15 (.050) -.10 (.202) .06 (.436) -.04 (.619) -.10 (.211) .04 (.561) .27 (< .001) 1.00 2 10 Speed 6.01 0.84 .04 (.577) -.16 (.018) -.06 (.383) -.05 (.474) -.10 (.138) -.05 (.442) -.10 (.161) -.08 (.227) .00 (.982) 1.00 1 11 Speed Variation 1.43 0.39 .09 (.212) .15 (.046) .39 (< .001) .01 (.877) .15 (.054) .07 (.390) .27 (< .001) .04 (.596) -.09 (.260) -.48 (< .001) 1.00 2 11 Speed Variation 4.04 0.96 .00 (.976) -.11 (.115) .12 (.089) -.03 (.651) -.04 (.602) .06 (.387) .10 (.138) -.04 (.547) .00 (.975) .53 (< .001) 1.00 1 12 Intra-Team Distance 9.59 5.59 -.05 (.468) .08 (.278) .18 (.017) -.03 (.659) .08 (.289) -.07 (.375) .06 (.392) .05 (.475) .09 (.245) .31 (< .001) -.01 (.850) 1.00 2 12 Intra-Team Distance 12.65 4.60 .08 (.261) .05 (.437) .06 (.346) -.03 (.645) .08 (.217) -.02 (.729) .01 (.833) -.01 (.845) -.04 (.522) .04 (.535) .17 (.011) 1.00 1 13 Route Deviation 7.47 4.06 -.12 (.115) .04 (.631) .02 (.759) -.03 (.728) .04 (.562) -.07 (.375) .11 (.151) .00 (.977) .14 (.058) .08 (.279) .10 (.197) .12 (.102) 1.00 2 13 Route Deviation 11.39 5.88 -.06 (.371) .02 (.745) .20 (.004) .06 (.354) .07 (.328) .03 (.647) .12 (.084) -.01 (.927) .11 (.122) .05 (.431) .01 (.877) .16 (.020) 1.00 1 14 Variation Route Deviation 3.06 1.97 -.04 (.582) .16 (.030) -.04 (.561) .01 (.944) .12 (.110) .06 (.441) .19 (.011) .15 (.046) .30 (< .001) .24 (.001) .05 (.490) .19 (.012) .58 (< .001) 1.00 2 14 Variation Route Deviation 4.77 2.39 -.11 (.120) -.01 (.836) .25 (< .001) .01 (.874) .00 (.965) -.04 (.596) .20 (.004) -.02 (.775) .06 (.398) .07 (.295) .15 (.027) .15 (.030) .62 (< .001) 1.00

Note. p-values less than .050 are in bold.

3

Regression Analysis

Per experiment and for each of the five GPS-based outcome variables we conducted regression analyses with the state constructs as predictors. We tested these ten regression models for random effects per participant, team, and Round. Testing for random effects is necessary since the measurements for each participant are not independent but depend on the Round that is measured and the team a participant is in. For instance, a team with fast walking members could have motivated a slow walking member to walk faster. To test the random effects, we created six random effect models which are displayed in Table 5 (see Model 1.1 – 6.1). The models are sorted by complexity with the lowest complexity in the beginning and the highest complexity at the end. The first model is a baseline model (Model 1.1) with a GPS outcome variable and 1 as predictor, and afterwards the random effects are added to the baseline model. Additionally, in Model 5.1 and 6.1 the predictor 1 is replaced with Round. For each Round, we assumed that the items varied together (covariance) and the time between measurements is equally spaced. Therefore, we used a first-order autoregressive covariance structure to model the covariance (Field, Miles, & Field, 2012).

To select a random effects model, a model fit indicator, such as the Akaike information criterion (AIC; Akaike, 1974), or the Bayesian information criterion (BIC; Schwarz, 1978), can be used. However, Barr et al. (2013) suggest it would preferable to select a random effects model based on the experimental design instead of selecting a model based on the model fit.

For our current study, a maximum random effects model included random slopes per round

and a static intercept per team and participant, and therefore Model 6.1 was selected. Finally,

for the current study, we were interested in the effect of all state variables on the GPS

outcome variable, and consequently, we added the State variables as predictors to the

regression model (Model 6.2).

Table 5

Random Effect Models

Model Random Intercept Random Slope Outcome variable Predictor

1.1 None GPS variable 1

2.1 Participant GPS variable 1

3.1 Team GPS variable 1

4.1 Team and participant GPS variable 1

5.1 Participant Round GPS variable Round

6.1 Team and participant Round GPS variable Round

6.2 Team and participant Round GPS variable Round and State variables

Model for Speed. We calculated Model 6.2 with Speed as outcome variable and Table 6 displays the results per estimate. As Table 6 highlights, Awareness Movement Change in Presence of Guards was a significant and positive predictor for Speed in Experiment 1 and the same relationship was not significant in Experiment 2 (b

= 0.08, p

< .001, b

= 0.05, p

= .241). Thus, when the participants reported a speed or route change after seeing the guards, the participants walked faster. An apparent explanation could be that the participants increased their speed, in an attempt, to outpace the guards.

Additionally, in Experiment 1 Suppressed Impulses to Change Movement is a

significant and negative predictor for Speed and the same relationship was not significant in

Experiment 2 (b

= -0.09, p

= .008, b

= -0.07, p

= .195). This means that, people which

suppressed their impulses walked slower, and an explanation could be that participants

walked slower in order not to attract the attention of the guards. An alternative explanation

could be that participants were uncertain which route would be the best to avoid the guards

and therefore slowed their pace.

Table 6

Regression Model for Speed: Statistics per Estimate

Experiment 1 Experiment 2

Estimate

Round 0.05 0.04 .232 0.11 0.07 .116

Illegal Card Selection -0.03 0.07 .636 0.04 0.18 .819

Alertness to Being Target of Guards -0.03 0.02 .166 -0.01 0.04 .721

Cognitive Self-Regulation -0.02 0.02 .395 0.01 0.05 .832

Situational Self Awareness -0.01 0.03 .687 -0.05 0.05 .339

Frightened by Presence of Guards 0.03 0.03 .290 -0.00 0.05 .951

Suppressed Impulses to Change Movement

0.03

-0.07 0.05 .195

Contemplation of Hostile Intent 0.02 0.03 .527 -0.03 0.05 .537

Awareness Movement Change in Presence of Guards

0.02

0.05 0.04 .241

Note. p-values less than .050 are in bold.

Model for Speed Variation. We calculated Model 6.2 with Speed Variation as outcome variable and Table 7 displays the results per estimate. The table shows that, Suppressed Impulses to Change Movement was a significant a positive predictor in Experiment 1 but not in Experiment 2 (b

= 0.12, p

< .001, b

= 0.03, p

= .575). This means, when the participants had suppressed impulses then they varied more in their walking pace. A simple explanation could be that participants failed in suppressing their impulses and therefore varied more.

However, as Table 6 shows, participants reduced their pace when they had suppressed

impulses (b

= -0.09, p

= .008, b

= -0.07, p

= .195) and if participants had failed in

suppressing their impulses, one would suspect that their overall pace increased and not

decreased. Hence, an alternative explanation could be that Suppressed Impulses to Change

Movement measures the uncertainty of the participants on how to avoid the guards and not

the suppressed impulses. Accordingly, the uncertainty could have caused the participants to

slowdown, in order to orient themselves, and then to follow the new path with an increased

pace.

Furthermore, Round is a positive and significant predictor for Speed Variation in Experiment 1 and the same relationship is not significant in Experiment 2 (b

= 0.08, p

< .031, b

= 0.05, p

= .575). Consequently, with each consecutive Round the participants varied more in their pace, and the variation could have helped the participants to avoid the guards better.

Additionally, Alertness to Being Target of Guards is a positive and significant predictor for Speed Variation, and the same relationship was not significant in Experiment 2 (b

= 0.05, p

= .008, b

= 0.07, p

= .139). Namely, when participants were targeted by the guards, then they would vary more in their speed. A likely explanation is that participants tried to avoid the guards by changing their walking pace.

Finally, Awareness Movement Change in Presence of Guards was a significant and

negative predictor for Speed Variation in Experiment 1 but not in Experiment 2 (b

= -0.06,

p

= .002, b

= 0.00, p

= .993). This means, when participants were aware that they changed

their route or speed after seeing the guards, then they varied less in their walking pace. A

reason could be, that participants had chosen a route, after seeing the guards, that avoided

the guards successfully and therefore the participants could keep their pace. Because of the

lower guard ratio in Experiment 1 it was easier to avoid the guards.

Table 7

Regression T, P and Beta Values for Speed Variation as Dependent Variable

Experiment 1 Experiment 2

Estimate

Round

0.04

0.05 0.09 .575

Illegal Card Selection 0.11 0.06 .088 -0.16 0.22 .449

Alertness to Being Target of Guards

0.02

0.07 0.04 .139

Cognitive Self-Regulation -0.03 0.02 .217 0.07 0.06 .285

Situational Self Awareness 0.04 0.02 .117 -0.11 0.06 .069

Frightened by Presence of Guards -0.03 0.03 .207 -0.01 0.06 .885 Suppressed Impulses to Change Movement

0.03

0.03 0.06 .575

Contemplation of Hostile Intent -0.05 0.03 .065 -0.04 0.06 .468

Awareness Movement Change in Presence of Guards

0.02

0.00 0.05 .933

Note. p-values less than .050 are in bold.

Model for Intra-Team Distance. We calculated Model 6.2 with Intra-Team Distance as outcome variable and Table 8 displays the results per estimate. The table shows that Round was a positive and significant predictor for Intra-Team Distance in Experiment 1 and the same relationship was close to significant in Experiment 2 (b

= 2.24, p

< .003, b

= 0.85, p

= .067).

This means that the distance to other team members increased with each Round, and the increasing distance could be a strategy, by the participants, that emerged to better avoid the guards.

Furthermore, Frightened by Presence of Guards was a significant and negative predictor for Intra-Team Distance in Experiment 1 and the same relationship was not significant in Experiment 2 (b

= -0.62, p

= .023, b

= -0.20, p

= .242). Therefore, when participants had feelings of fright because of the guards then they walked closer together, possible to compensate for the fear.

Additionally, Contemplation of Hostile Intent was a significant and positive predictor

for Intra-Team Distance in Experiment 1 and the same relationship was not significant in

Experiment 2 (b

= 0.52, p

= .034, b

= 0.09, p

= .580). This means, when participants were